PPL: Adaptive Message Passing Interface (AMPI)

Adaptive Message Passing Interface (AMPI)

Parallel Computational Science and Engineering (CSE) applications often exhibit irregular structure and dynamic load patterns. Many such applications have been developed using procedural languages (e.g. Fortran) in message passing parallel programming paradigm (e.g. MPI) for distributed memory machines. Incorporating dynamic load balancing techniques at the application-level involves significant changes to the design and structure of applications. On the other hand, traditional run-time systems for MPI do not support dynamic load balancing. Object-based parallel programming languages, such as Charm++ support efficient dynamic load balancing using object migration for irregular and dynamic applications, as well as to deal with external factors that cause load imbalance. However, converting legacy MPI applications to such object-based paradigms is cumbersome. AMPI is an implementation of MPI that supports dynamic load balancing and multithreading for MPI applications.

Our approach and implementation is based on the user-level migrating threads and load balancing capabilities provided by the Charm++ framework. Conversion from legacy codes to this platform is straightforward even for large legacy codes. We have converted the component codes ROCFLO and ROCSOLID of a Rocket Simulation application to AMPI. Our experience shows that with a minimal overhead and effort, one can incorporate dynamic load balancing capabilities in legacy Fortran-MPI codes. An introductory tutorial about how to convert existing Fortran 90 codes to AMPI by hand can be found here.

Also, please refer to the talk on AMPI and AMPI Tutorial.

Software

AMPI has been integrated into the Charm++/Converse distribution. Please download the source distribution. And install AMPI by specifying the target "AMPI" to the Charm++ build command. (For details, see the README file in the source distribution.) Also, see the Adaptive MPI Manual. [postscript] [PDF] [html]

People

Papers

07-08 Chao Huang, Gengbin Zheng, Laxmikant V. Kale, Supporting Adaptivity in MPI for Dynamic Parallel Applications, PPL Tech Report 07-08
05-04 Chao Huang, Gengbin Zheng, Sameer Kumar, Laxmikant V. Kale, Performance Evaluation of Adaptive MPI, Proceedings of ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming 2006
03-07 Chao Huang, Orion Lawlor, L. V. Kale, Adaptive MPI, Proceedings of the 16th International Workshop on Languages and Compilers for Parallel Computing (LCPC 2003), LNCS 2958, pg 306-322
00-03 Milind Bhandarkar, L. V. Kale, Eric de Sturler, and Jay Hoeflinger, Object-Based Adaptive Load Balancing for MPI Programs, In Vassil Alexandrov et al (Eds), Computational Science -- ICCS2001, Proceedings of the International Conference on Computational Science, San Francisco, CA, Lecture Notes in Computer Science, Vol. 2074, Springer Ver;ag, pp 108-117, May 2001.